Harvard Magazine
Main Menu · Search ·Current Issue ·Contact ·Archives ·Centennial ·Letters to the Editor ·FAQs

Back. Also see the sidebars The "Engine Man's" Dream Journal, From Neurobiology to Psychology, and A Degree Will Not Be Forthcoming.
These small charts show the firing of individual neurons (upper bar) and either eye movement or electroencephalogram (EEG) data (lower lines).

Hobson made this connection between his neurobiology work and dreaming in 1973, just as he was about to address a scientific conference in Scotland. "So," he recalls, "I decided to present the concept to the conference in a lecture entitled 'The Brain as a Dream Machine,' which attracted attention and provoked intense debate."

Part of the debate centered on the source of Hobson's data. He and his colleagues had done virtually all of their experimental work on cats. As noted, while the REM sleep of cats seems quite similar to that of humans, there was no obvious way for Hobson to support his theoretical arguments by eliciting dream reports from cats or assessing the nature of cat dreams.

But the literature of human psychology is full of dream reports. So Hobson and McCarley decided to take collections of human dream reports, give them an old mental-health test known as the "mental status examination," and see if the results fit their theory of brainstem regulation.

Hobson and McCarley realized something else as they entered the world of dream psychology. They knew their dream theory clashed with Freud's old belief that dreams seem strange only because they are "cut" by a hidden censor. As Hobson wrote later, "I differ from Freud in that I think that most dreams are neither obscure nor bowdlerized, but rather transparent and unedited. They contain clearly meaningful, undisguised, highly conflictual impulses worthy of note by the dreamer (and any interpretive assistant). My position echoes Jung's notion of dreams as transparently meaningful...."

As this indicates, the new dream concept, which Hobson and McCarley called the "activation-synthesis hypothesis," was not the first to oppose Freud's. But it was the first one firmly grounded in neurobiology. On the down side, the theory seemed unlikely to prove either popular or saleable, because it rejected all symbolic dream interpretation schemes out of hand. Furthermore, it seemed likely to suffer benign neglect from psychologists because it was not yet firmly grounded in psychology, was unfamiliar to the psychoanalytic community, and directly contradicted Freud. So the real danger, as Hobson and McCarley saw it, was not that their theory would be rejected, but that it would be ignored.

To keep this from happening, Hobson and McCarley wrote two rigorous articles for the American Journal of Psychiatry in 1977 dismissing Freud's old dream theory and offering activation-synthesis as a replacement. For flavoring, Hobson shaped the articles into an overt assault on the foundations of psychoanalysis. The result was encouraging. The pieces generated more letters to the editor, mostly from outraged psychoanalysts, than any articles previously published in the Journal. "I would admit to having created some heat where light might have been more useful," said Hobson later, "but I can tell you, they weren't paying any attention until I turned the heat up a bit."

Hobson kept the juices flowing that year by masterminding an art/science exhibit called Dreamstage: A Multimedia Portrait of the Sleeping Brain. The main feature of this exhibit, which ran six weeks at Harvard's Carpenter Center, was a person sleeping in a closed room behind a one-way mirror, in plain view of an audience sitting in a darkened room on the other side of the mirror. As in a scientific sleep lab, the sleeper was hooked up to instruments that monitored brain waves, eye movements, and body movements. But instead of applying their data to paper charts, the instruments used it to paint colored waves of laser light on the walls and to run a synthesizer that converted the sleep data into music. When the sleeper shifted or started dreaming, the patterns of light and sound would change, and people would rush in to find out what was happening. The event drew more than 10,000 viewers, received nationwide media coverage, spawned a six-city road show, and served its purpose of announcing to the world that science was starting to learn the truth about sleep and dreams.

Of course, the two Hobson-McCarley theories and Dreamstage had their share of critics. Not just psychoanalysts, but a wide range of natural scientists and psychiatrists had reservations. This was only natural, because science demands that new theories be critiqued and tested; sharp disagreement typifies the cognitive sciences; and certain details of the two theories required adjustment. Beyond that, some scientists tended to focus on details to a point where they rejected these broad theories out of hand and responded to Hobson's open advocacy with name-calling. As one Hobson defender put it, "Allan's an unusual scientist, in that he does propose fairly general theories. I think people who say he's a flimflam artist are the ones who aren't smart enough to do that kind of thing themselves."

But what really exposed Hobson to professional criticism was Dreamstage. Like medieval monks, all scientists take an unspoken vow not to pander their wares to laymen. If they break that vow, like former Harvard and Cornell astronomer Carl Sagan, they risk paying for public acclaim with professional discredit. Perhaps because Dreamstage was really just one singular happening, Hobson escaped this fate. Nevertheless, when he was granted tenure around this time, one member of the committee making the decision told him, "We decided to grant you tenure despite Dreamstage."

Hindsight shows this to have been a good move for Harvard, because the Hobson-McCarley theories proved robust. Not only did the underlying research survive critical review, but no other broad theories based on neurobiology arose to take their place. As Richard Davidson, a leading experimental psychologist at the University of Wisconsin, explains, "By and large, these theories have held up well. Some of the specific details have been revised in the light of modern data. But particularly the activation-synthesis theory of dreaming has provided a framework for relating neurobiologic and mental activity, one that is still a guiding model in all the modern efforts to understand the underlying functional significance of REM sleep and dreaming."

Having thus built a bridge between neuroscience and the mind, Hobson set out to reinforce it. He began writing The Dreaming Brain, finished and published in 1988, which remains among the best of the scientific books on dreams. While continuing his animal work, he pursued research
on human sleep and dreams with whatever tools he had available--including the Dreamstage sleep lab when Dreamstage was in progress. He used the old Dreamstage equipment, once Dreamstage was done, to set up a sleep lab at the Massachusetts Mental Health Center. And he invented a simple sleep-charting device, picturesquely dubbed the "Nightcap," to monitor REM sleep and substitute for the relatively costly and cumbersome sleep lab.

This "Nightcap" idea grew out of time-lapse photos that Theodore Spagna, a principal Dreamstage collaborator, took of the exhibit's sleepers. These pictures suggested to Hobson, as other evidence confirmed, that normal sleepers are likely to change position only when entering or leaving REM sleep. We now know that's because in nonREM sleep the brain is not usually excited enough to cause such movement, while in REM sleep the motor commands that stream toward the spinal cord are blocked, paralyzing the sleeper. It is therefore just at the beginning of REM (before the block is in place), and just at the end (as the block is deactivated), that motor commands from the newly excited or still excited brain are likely to cause a sleeper to shift about.

The Nightcap, which took advantage of this discovery, consisted of a simple piece of headgear wired to a smart pocket-sized recorder. The device reliably detected REM sleep by recording eye movements (indicating REM sleep) plus head movements (indicating the beginning or end of REM sleep and so providing a check on the eye data). At some point the Nightcap's wearer, awakened by an alarm or naturally, would report any dream recalled. In this way, like the sleep lab, the Nightcap was able to track a volunteer's sleep patterns, provide a dream report, and indicate whether the last sleep stage (in which the dream occurred) was REM or nonREM sleep.

This inexpensive and versatile device soon proved its worth. Hobson used it at volunteers' homes to collect masses of dream reports far exceeding any previous collections, with the sleep stage of each report indicated; and other investigators wishing to track sleep patterns found it useful in a wide array of diverse settings where sleep labs couldn't go.

Around the time the Nightcap was emerging, McCarley left Hobson's lab to head up a research facility at the Harvard-affiliated Brockton Veterans' Administration Hospital. (Now professor of psychiatry and head of the hospital's psychiatry department, he is still actively pursuing sleep research, focusing especially on the causes and chemistry of sleep.) Soon there were other changes. Hobson began offering a course at the Extension School (Psychology E1450, "The Biopsychology of Waking, Sleeping, and Dreaming") that he still teaches. He got a MacArthur Foundation Mind-Body Network grant for studying human sleep and dreaming that permitted him to expand the human-research side of his lab's work. And he used his own infectious enthusiasm to recruit volunteers from the extension course and elsewhere. "We've always been very open door," Hobson says. "People hear about us, they like the ideas in the first place, and we're very friendly. So we have extension school students coming, we have Harvard College students coming...."

After 1977, much of the Hobson laboratory's dream work focused on strangeness--the odd discontinuities, mismatching elements, or uncertainties that nearly all REM dreams contain. Clearly, these bizarre features accounted for much of the chaotic and choppy nature of REM dreams. Equally clearly, the strangeness in certain cases seemed related to basic sensory or motor problems. For instance, when you dream you are running but getting nowhere, it seems likely that your brain's motor centers (active in REM sleep) are issuing commands to run; but since these commands are blocked by the brainstem and you don't move, your brain perceives running without motion. Likewise, when you dream you are falling endlessly, it probably means your brain is receiving no sensory clues to indicate your position. Since your brain always receives such clues except when you are in REM sleep or falling freely with your eyes closed, your brain concludes that you are falling.

In seeking to explain such dream features, Hobson had tended to emphasize the intimate connection between brainstem REM neurons and sensory neurons transmitting information about the body's position. He and his people went to work on such ideas by examining masses of dream reports, classifying their bizarre features, and asking whether these features appeared due to sensorimotor inputs, brain censorship, or something else. They found, as Hobson recounts in The Dreaming Brain, that "the trouble arises at or near the sensorimotor level," with most of the strangeness deriving either from bizarre sensorimotor signals or from the brain's efforts to make sense of them. By and large, the findings support the activation-synthesis idea.

Similarly, Hobson's group sought to track REM dream emotions to their neurobiologic roots. Among other things, thoroughgoing examination of dream reports showed that even though emotions were not referred to in all cases, those mentioned were frequently strong and unpleasant. Says Hobson: "Anxiety (most common), elation (next), and anger (next) accounted for about 70 percent of the emotions cited. After that the numbers drop way down, with the rates for shame, guilt, affection, and erotic feelings being very low, all under 5 percent." These results are consistent with activation-synthesis and also with excitement of the brain's prime emotion center, an almond-shaped body called the amygdala, that is known to receive strong stimulation from the brainstem in REM sleep.

While this work was advancing, Hobson had not forgotten that his laboratory's greatest strength was its ability to track mental events all the way from the actions of individual neurons to the broadest reaches of theoretical psychology. Therefore, he continued to direct animal research--work now spearheaded by two assistant professors at the medical school, James Quattrochi and Bernat Kocsis, who are, respectively, studying brain anatomy and key electrophysiologic interactions of certain brain substructures. At the same time, he came to work closely with a number of psychology researchers--among them Edward Pace-Schott and Robert Stickgold '66, both recruited through the extension course. Pace-Schott, an instructor in psychiatry, now oversees Nightcap work. Stickgold, an assistant professor of psychiatry at the medical school, is co-teaching the extension course with Hobson and pursuing research on how sleep affects the mind.

Recently, some of the Hobson-Stickgold work has shown real promise. In one experiment, dream reports containing discontinuities were selected. Half the reports were left intact, while the other half were cut apart and recombined. Judges were then asked to single out the spliced reports.

As Hobson relates with pleasure, "They couldn't do it. I was sure, by the way, that I would be able to do this when it was suggested by Bob Stickgold in a seminar. It turns out you can do it if and only if the splices occur in the midst of continuous action. Otherwise it's hopeless. If you take it scene by scene, there is no objective continuity, even though all who read such reports are convinced it must be there.

"Everyone, including the content-analysis people, have always assumed that dreams are stories, or at least books, or chapters, or something. This work shows they aren't. Indeed, we concluded that most of the integrity that is seen in these reports is projected onto them by the viewer. It's as if what's going on is so difficult to integrate that it just can't be done, and so you get one little story, and then another and another.

"This has profound importance, not just for dream theory but for how people perceive things (how they read, for instance) and how they understand the world. The brain is always trying to make a whole story out of whatever facts are presented. That's the brain's job. You know you couldn't function in the world unless you did that. But with dreams it's a fool's errand. In other words, there is probably a lot more discontinuity and incongruity than we allow ourselves to admit, not just in dreams but in the world in general, because we just can't stand it, we have to be holistic."

If the purpose of dreams is not to release libidinal id, as Freud maintained, then what is it? Hobson and many others think it could relate to reorganizing the mind's content and reordering memories. "This is an important shift in emphasis," he says, "one that allows dreams to be nonsensical while the task being done is crucial. The scrambling isn't being done to hide forbidden thoughts, it's being done for some other purpose, and that purpose could be the reordering of memory and the reinforcing of links between memory and emotion.

"What you probably can't have is a 'penny scale' of dream interpretation, a code which says 'This means that.' But you can have so much more. Your brain is doing all this homework for you, getting your mind in order. In this light, dreams may be nonsensical for the same reason that housecleaning tends to make a mess. My housekeeper makes a mess every time she visits, but it's much cleaner when she's done."

Pursuing this line of reasoning, Hobson's team has been investigating how sleep relates to memory and learning. Researchers elsewhere had found that volunteers could "learn" subliminally. They did so by showing their volunteers screens on which a pattern of short lines was flashed briefly. Most of the lines were horizontal, but three were diagonal lines arranged either side by side or one above another. As the patterns were displayed for shorter and shorter times, the volunteers found it progressively harder to tell whether the diagonal lines were arranged vertically or horizontally, and eventually they couldn't tell at all. However, if asked to repeat the tests later, they would typically do better on the second trial--but only if several hours had passed since they first took the test. The conclusion: learning was taking place and being consolidated very slowly, outside of conscious awareness.

Hobson team members, including Stickgold and Vitul Patel '98, decided to test sleep-lab volunteers to see how such learning relates to sleep patterns. So far their work suggests that long, deep sleep ending with a lot of REM sleep helps consolidate such learning, while failure to get proper sleep can wipe it out. This tends to reinforce the long-held suspicions of innumerable undergraduates--that going short on sleep to cram for tests may help pass the tests, but the information "learned" is soon gone. Of course, late-night cramming may provide good general training for the brain, and to date the Hobson team's research appears to be dealing with only one specialized kind of learning. Even so, this work could wind up giving new meaning to the old dictum that education is what's left after you've forgotten everything you learned in school.

Meanwhile, Edward Pace-Schott is coordinating an array of Nightcap projects with other institutions, using the device to monitor the sleep patterns of many groups--including volunteers taking antidepressant drugs, chronic insomniacs receiving anti-insomnia behavior therapy, truck drivers slumbering in their cabs, and astronauts and cosmonauts sleeping on the Mir space station. Surprisingly, as Stickgold recently pointed out, "The Nightcaps on the Mir were some of the only scientific research gear to survive the collision between a supply capsule and the main Mir module housing most of the Mir's scientific experiments. The Nightcaps were not in this module at the time, and so were saved. They are currently still working."

Hobson's sleep lab, containing much of the technical gear inherited from Dreamstage, is also still working, because it is needed for tasks like testing volunteers immediately after waking. For most purposes, however, the Nightcap has tended to supplant it. In a like manner, while animal work is still needed for certain kinds of brain research, neurobiology's horizons are being broadened by new scanning techniques, notably PET (positron emission tomography) and MRI (magnetic resonance imaging), that permit researchers to observe the internal workings of a volunteer's brain under controlled conditions. Although Hobson is not directly involved in much of this scanning activity, he is elated because "the work is starting to confirm--with pictures of the human brain--what could previously be surmised only from animal research."

Despite its astonishing progress, the sleep and dream revolution is far from over. Many researchers now believe sleep's main functions are to rejuvenate the brain and consolidate memory in various ways, but no one has proven that. Likewise, many think dreams are incidental side-effects of this process that evolution causes us to forget because the forgetting does no harm, while remembering might confuse dreams with reality, threaten our mental balance, and adversely affect survival. But again, nobody knows for sure.

Paradoxically, none of the spectacular advances made to date have much daunted Freudian psychoanalysts or other dream interpreters. What has dethroned psychoanalysis in recent years is an array of powerful psychoactive drugs effective in treating mental illness, most of which mimic or influence the same brainstem chemicals that control REM sleep and dreaming. Meanwhile, humanists' regard for Freud and his dream theory has been only slightly dampened, while public enthusiasm for a wide range of ad hoc dream interpreters is on the rise.

In hindsight, the reason seems clear. People love both mystery and knowledge. Astrology did not vanish when Copernicus and others started learning the truth about the planets; instead, that was merely the place in the road where astronomy and astrology diverged. Similarly, the great strength of Freud's dream theory arose from its appeal to both mystery and reason. That appeal was needed in Freud's day, because the facts could not be known, just as such an appeal is impossible today because we know too much.

Thus, like modern Copernicans, Hobson and other scientific dream-catchers of recent times have passed the place where the road divides, and they cannot go back. They will never rekindle the enthusiasm lit by Freud's dream theory, because they cannot legitimately make his dual appeal. Instead, like Adam after he first bit the Apple of Knowledge, they must settle for the lesser role and lesser fame of those who merely succeed in dispelling ignorance by discovering the truth.

Jonathan Leonard '63, the former editor of an international public-health journal, is a professional writer specializing in medicine and science.

Main Menu · Search ·Current Issue ·Contact ·Archives ·Centennial ·Letters to the Editor ·FAQs
Harvard Magazine